01
710 curated coding exercises
Work through 410 Python, 150 Java, and 150 SQL drills across Easy, Medium, and Hard levels.
Other projects
KitCode Practice Studio
Practise Python, Java, SQL, and machine-learning interview concepts in one focused Windows workspace. KitCode combines a curated exercise bank, local code execution, progress tracking, reviewed answers, and optional live coaching from Codex—without requiring a Git or terminal workflow.
Free download from GitHub · Windows · core practice works offline after setup
710
curated Python, Java, and SQL coding exercises
400
interview and machine-learning concept questions
3
hands-on coding tracks with Easy, Medium, and Hard levels
The workspace
Practice that feels closer to building real software.
KitCode works well for first-time practice, interview preparation, refreshing fundamentals, learning data structures and algorithms, or building confidence before moving into larger projects.
Why learn with KitCode?
A complete, focused feedback loop.
The workspace brings the exercise, editor, runner, feedback, and progress together instead of spreading practice across several tools.
02
Live help from Codex
Ask questions naturally, request a targeted hint, or get an explanation based on the exercise and current code.
03
A complete feedback loop
Run code, inspect output, submit against the exercise test suite, and track completed problems.
04
Reviewed best answers
After a correct submission, compare the approach with an expected solution, Big-O target, and readability-first rubric.
05
Python line-by-line tracing
Step through executed lines and inspect changing values when a program is difficult to reason about.
06
Interview FAQ banks
Review 100 questions each for Python, Java, SQL, and machine-learning concepts without opening the code editor.
07
Local-first practice
The curated bank, runner, drafts, and progress stay on the computer and continue to work offline after setup.
How it works
A practical learning loop.
Work through an exercise
- Choose Python, Java, or SQL and select an Easy, Medium, or Hard practice level.
- Read the task, examples, constraints, hints, and expected complexity.
- Write the answer in the built-in Monaco editor.
- Press F5 to run it. Python exercises can also be traced line by line.
- Press F6 to submit it against the full exercise test suite.
- After passing, open Best answer to compare complexity, clarity, and implementation choices.
The selected level, drafts, solved exercises, FAQ scores, and recent coach conversations are remembered locally.
Optional live coaching
KitCode can connect to an authenticated Codex CLI and use it as a live programming coach. OpenAI, Anthropic/Claude, and compatible local models are also supported.
- Ask follow-up questions in a streaming conversation while you work.
- Request a concise, read-only hint anchored near the relevant editor line.
- Keep separate conversations for each exercise and AI provider.
- Retry or restart an interrupted response without retyping the question.
- Ask for explanations of errors, concepts, trade-offs, or possible next steps.
The coach receives the visible exercise context, current script,
cursor position, question, and a small amount of recent conversation.
Hidden tests and reviewed reference answers are never included.
Curriculum
Learning topics available.
| Track | Curated practice | Topics include |
|---|---|---|
| Python | 410 exercises: 176 Easy, 140 Medium, 94 Hard | Typed functions and dataclasses, iterators and generators, async I/O, retries, data structures, classes and OOP, searching, sorting, trees, graphs, dynamic programming, and advanced algorithms. |
| Java | 150 exercises: 35 Easy, 86 Medium, 29 Hard | Core syntax, arrays, strings, collections, maps, stacks, linked lists, binary search, trees, graphs, greedy algorithms, dynamic programming, backtracking, and design. |
| SQL | 150 exercises: 46 Easy, 89 Medium, 15 Hard | Filtering, aggregation, joins, NULL, CASE, subqueries, window functions, recursive CTEs, time-series analysis, gaps and islands, hierarchies, and reachability. |
| Python Fundamentals | 100 optional multiple-choice questions | Typed functions and classes, iterators, async I/O, exceptions, testing, dependency injection, API pagination, rate limiting, and clear abstractions. |
| Java Interview FAQs | 100 optional multiple-choice questions | Java semantics, OOP, exceptions, generics, collections, streams, concurrency, modern Java, testing, and the JVM. |
| SQL Interview FAQs | 100 optional multiple-choice questions | Filtering, joins, aggregation, window functions, CTEs, schema design, indexes, performance, transactions, security, and warehousing. |
| Machine Learning Concepts | 100 optional multiple-choice questions | Regression, gradient descent, preprocessing, regularisation, evaluation, forests and boosting, neural networks, transformer attention, and production ML. |
FAQ answers are checked locally with explanations, topic filters,
next-unanswered and review-misses views. SQL practice supports SQLite,
PostgreSQL, MySQL, and SQL Server syntax, executed safely against fresh
local SQLite fixtures.
More useful features
Small details that make practice easier.
- Searchable exercise banks and solved counts for each difficulty level.
- Automatic draft saving and recovery if a newer edit would otherwise hide earlier work.
- Public feedback plus private checks that discourage hard-coded answers.
- Direct class-API testing for Python OOP exercises.
- Optional AI-generated drills, clearly marked as provisional and kept separate from the curated 710.
- One-click hints from Kit, the movable desktop mascot.
- Keyboard-first Run and Submit shortcuts, responsive drawers, and accessible dialogs.
Built as a local full-stack application
The interface and exercise runner are packaged into a Windows download, while tagged releases are tested and published through GitHub Actions.
React
TypeScript
Vite
Monaco Editor
Python
FastAPI
SQLite
GitHub Actions
Try the project
Download KitCode from GitHub.
- Open KitCode-Windows.zip and select Extract All.
- Open the extracted KitCode folder.
- Double-click launch.bat.
KitCode opens in the browser. The first launch may need a few minutes and an internet connection while it prepares Python and the required packages; later launches are much faster. Java is prepared automatically where the computer permits it.